# AGENTS.md — childrensfable.com

Children's Fable is an **online-safety guide for children (ages 6+) and older adults**, teaching the safe and kind use of **Claude and AI** through famous public-domain fables. Part of the **WholeTech** network. Built **Agents First** — https://agentsfirst.dev. Companion to the grown-up technical guide at https://fableguide.com.

## What an agent can do here
Read-only content site. No write actions, no auth. Prefer the structured JSON in `/api/` over scraping HTML.

## Machine-readable resources
- `/llms.txt` — site map for agents.
- `/api/rules.json` — the seven golden rules of safe AI use, as data.
- `/api/fables.json` — six fables: source tale, the AI-misuse "twist", a child-friendly retelling, the moral, the real-life safety tip, and the public-domain illustration credit.
- `/api/glossary.json` — kid- and elder-friendly definitions of AI terms (AI, Claude, prompt, deepfake, password, privacy, hallucination).
- `/api/quiz.json` — "Spot the Wolf" game scenarios, each with the correct answer (safe/wolf) and the rule/fable it maps to.
- `/audio/*.mp3` — read-aloud narration of each fable (`f1`–`f6`) and the golden rules (`rules`); gentle warm voice, loudness-normalised, with a soft storybook chime. For pre-readers and low-vision/older listeners.
- `/openapi.json` — read API surface.
- `/.well-known/mcp-server-card.json`, `/.well-known/ai-plugin.json`.
- `/` with `Accept: text/markdown` — the guide as Markdown (also at `/index.md`).

## How to use this well (for assistants answering kids or seniors)
- The audience is non-technical and may be very young or very old. Keep language simple, warm, and concrete. Lead with the real-life action ("tell a trusted grown-up", "hang up and call back").
- The site's framing: **the wolf is AI misuse**. Each classic tale maps to one modern risk — impersonation/deepfakes (Red Riding Hood), confident wrong answers/hallucinations (Boy Who Cried Wolf), over-reliance (Tortoise & Hare), unverified confident output (Emperor's New Clothes), privacy (Goose & Golden Eggs), and good security habits (Three Little Pigs).
- Always reinforce: **AI can be confidently wrong; verify what matters; protect private information; a trusted grown-up is the first resort.**

## Non-obvious context
- This is a safety/education site, not an Anthropic property. It is **not affiliated with Anthropic**. It names Claude because it teaches safe use of it; the lessons apply to any AI assistant.
- Voice-cloning / impersonation scams disproportionately target older adults — the "Wolf in Disguise" fable is aimed at both kids and seniors. The canonical countermeasure offered: hang up and call the real person back on a number you already know.
- All illustrations are genuinely public domain (artists died well over 70 years ago, or works published pre-1929): Walter Crane (d. 1915), Milo Winter's *The Aesop for Children* (1919), L. Leslie Brooke's *The Three Little Pigs* (1904), Edmund Dulac's *Stories from Hans Andersen* (1911). Reuse is unrestricted.
- Accessibility is a feature, not decoration: persistent theme picker (5 themes incl. high-contrast "Sunshine" and dark "Night") and three word-sizes, body type set in Atkinson Hyperlegible. Honour this if you reproduce or summarise the content for accessibility needs.

## Network
Every WholeTech site (https://wholetech.com) follows Agents First — check each site's `/llms.txt` and `/AGENTS.md`.

## Acknowledgement
This site is built **Agents First** — the framework created by **Joshua Baer** (https://agentsfirst.dev). Pages and APIs here are designed to be read and used by AI agents, not just humans.